Commercial thinning (CT) aims to mitigate inter-tree competition, allow immediate timber recovery, and improve the growth, value, and resilience of residual trees. However, the implementation and assessment of CT is often complicated by costly field plots that may not capture the heterogeneity of pre-treatment and post-treatment CT stands. Field plots may also struggle to quantify CT responses, particularly when sample sizes are small, repeated measurements are not possible, and unthinned controls are not available. To address these issues, we propose a spatial CT assessment framework that employs multitemporal airborne laser scanning, individual tree detection, and a spatially-explicit growth model to quantify CT removals, project long-term CT responses, and compare yield outcomes between thinned and unthinned scenarios. Then, we apply this framework to eight CT stands in British Columbia, Canada (295 ha total; 366,863 lidar-detected trees), model these stands from initial stand ages of 31-40 to age 100, and explore CT impacts on growth, competition, merchantable volume, and piece size. CT treatment intensity was not uniform and varied spatially between and within stands, with mean basal area removals ranging from 23.9% to 42.6%. Under long-term projections, thinned stands yielded 49% more large-log volume (>40 cm diameter) at age 80 (195 vs 131 m(3) ha(-1)), and unthinned stands maintained higher overall stem volumes with fewer large trees. After accounting for CT removals, cumulative volume growth was similar between thinned and unthinned scenarios. This framework shows promise to support the establishment, monitoring, and enhancement of CT treatments in operational stands.

The conventional growth models used in forest management often rely on historic biometric relationships and do not consider climate’s impact on growth. Climate sensitive predictions of forest growth are essential to assess sustainable forest management and forest carbon, particularly under increasing climate change. In this study, we explored volume and stem biomass predictions from the climate sensitive, hybrid/process-based growth model 3-PG (Physiological Principles in Predicting Growth) for four tree species in British Columbia, Canada. Then, we used 3-PG to climate-adjust volume predictions from a conventional growth model without climate sensitivity. Yields from 3-PG and this model fusion were evaluated using repeated measurement plots. Stem biomass and volume predictions from 3-PG tracked the observed data, producing Relative Model Biases (RMBs) between 1 and -8

Wildfire fuel-reduction treatments in British Columbia (BC) generate large volumes of underutilized biomass. This study assesses the technical suitability and supply chain economics of using biomass from thinning and surface fuel reduction for small-scale combined heat and power (CHP) systems. Biomass availability, including non-delimbed and delimbed energy wood, residues, and surface woody debris, was evaluated in a 12-ha forest stand in the interior of BC, Canada. Woodchip quality was analyzed for particle size distribution, ash content, moisture content, and bulk density. Supply chain costs were estimated for multiple recovery methods, machinery scales, and chipping locations, considering specific treatment options. Results showed that one year of natural air drying enabled non-delimbed and delimbed energy wood to meet CHP system requirements for particle size and ash content, whereas residues and woody debris required targeted post-processing. Supply chain costs ranged from $175-426 per oven-dry tonne, driven primarily by transportation distance and treatment intensity. These findings demonstrate that fuel-reduction biomass can support small-scale CHP system operations when feedstock is locally sourced and of sufficient quality. Integrating energy production with higher-value product streams may provide a practical approach to cost-effective biomass utilization while contributing to wildfire risk mitigation, local energy security, and the regional bioeconomy. This study offers a transferable framework for evaluating operational feasibility and supply chain design in other fire-prone regions.

In times of unprecedented global change, forest management education, especially in silviculture, must evolve to prepare future forest managers with relevant skills and a comprehensive view of adaptive silviculture. As a data-driven science, silviculture must now integrate multidisciplinary technical and professional expertise to shape the forests of tomorrow—by planning and assessing treatment impacts while also meeting socioeconomic needs. As a result, silviculture education, particularly in undergraduate programs, needs to advance to meet these future challenges to ensure students have the essential tools and analytical skills required to undertake holistic, field-based silviculture practice. To support this advancement, we propose a vision for silviculture education that emphasizes fundamental knowledge (e.g., forest ecology, mensuration, governance), while also adapting to evolving concepts driven by socioeconomic factors, new silvicultural systems, a focus on ecosystem services, and the availability of new geospatial technologies. This new vision calls for strong leadership in experiential education to foster active learning through innovative tools, work-integrated experiences, and interdisciplinary collaboration. Such a curriculum will equip students with the skills required to integrate knowledge and adapt holistically, preparing them to meet the evolving demands of modern silviculture.

1. Exacerbated by climate change, wildfires in British Columbia, Canada, have increased in extent and severity, impacting forests, including commercially valuable species like interior Douglas-fir (Pseudotsuga menziesii var. glauca). Post-wildfire salvage logging aims to mitigate financial losses and accelerate regeneration, though its ecological impacts remain uncertain. 2. This study was conducted in the Alex Fraser Research Forest, where a 2017 wildfire burned approximately 1000 ha. Combined with 2023 seedling biomass and %N measurements, we used linear mixed-effects models to examine the physiological responses of regenerating interior Douglas-fir seedlings to burn severity and salvage logging, using carbon, nitrogen, and oxygen stable isotope analyses (S13C, S15N, and S18O) to assess water-use efficiency (WUEi), photosynthesis, water stress, and nitrogen cycling post-disturbance. 3. Higher seedling biomass was found in high-severity, not-salvaged sites. Moderate-severity, not-salvaged sites had lower S13C and S18O values compared to high-severity (salvaged and not-salvaged) and moderates-severity, salvaged sites. Higher leaf %N was positively correlated with S13C values across treatments, indicating enhanced water-use efficiency. 4. The statistically significant interactions between burn severity and salvage logging and their influence on seedling biomass, S13C, and S18O emphasize the key role of microclimatic conditions in post-fire recovery. In high-severity sites, salvage logging did not enhance seedling biomass, likely due to already sufficient light availability. In moderate-severity sites, salvage logging had small, positive effects on seedling biomass that were not statistically significant. Higher leaf nitrogen content appeared to boost WUEi across treatments. These findings support tailoring post-wildfire management to burn severity, with minimal intervention in high-severity areas and selective salvage in moderate-severity sites.

ABSTRACT Wildfires are increasing in frequency and severity due to climate change, posing challenges to forest ecosystems, including the southern interior of British Columbia, Canada. Interior Douglas‐fir (Pseudotsuga menziesii var. glauca) is a species of great cultural, ecological, and economic importance, necessitating the investigation of post‐wildfire regeneration amidst this changing wildfire regime. This study examines interior Douglas‐fir seedling regeneration across three burn severity levels (low, moderate, high) 5 years post‐wildfire at a site in interior British Columbia. Natural regeneration and seedling traits were measured in 2022 and paired with stable isotope analyses (δ13C, δ15N, δ18O) and foliar nutrient assessments. We employed linear mixed‐effects models to assess the impact of burn severity and light, water, and nutrient factors on seedling biomass. Results indicate higher seedling density in low severity sites but larger individual biomass in moderate and high severity sites. Light availability was the primary factor limiting individual seedling biomass, with greater δ13C and biomass in high severity sites, suggesting that reduced canopy cover enhances photosynthesis and water use efficiency. Despite higher solar exposure, seedlings in high severity sites did not show increased drought stress according to leaf δ18O and stem water contents, likely due to reduced interception and competition for soil water by overstory trees. Biomass growth was not linked to leaf nutrient status, indicating nutrient availability, particularly N, did not limit seedling biomass. While light availability is the current primary growth‐limiting factor for regenerating interior Douglas‐fir seedlings in this study, increased frequency and intensity of heat waves and droughts associated with climate change may increase water stress, emphasizing the need for long‐term monitoring and adaptive management to support the regeneration of interior Douglas‐fir.

Accurate tree species mapping is essential for effective forest management but is often constrained by manual, labour-intensive workflows that limit scalability. While airborne laser scanning (ALS) supports large-scale forest attribute prediction, species classification remains difficult in complex, multi-species forests. To address this, we propose an automated, data-driven dual-stream deep learning framework that integrates ALS data with pointcloud metrics to identify individual tree species. Our framework incorporates an automated approach to individual tree segmentation and species labelling using existing forest inventory and field data, resulting in a dataset of 16,269 labelled individual tree point-clouds of four species across a 630,000 ha boreal mixed species forest in Ontario, Canada. Our dual-stream deep learning model integrates a Point Extractor to generate feature representations from raw ALS point-clouds and a complementary Metrics Network to process the point-cloud metrics. Results, based on the split test set of 2441 trees, showed that the inclusion of the Metrics Network improved tree species classification accuracy by approximately 11 % compared to models that rely solely on the Point Extractor. A weighted F1-score of 0.70 and area under the receiver operating characteristic curve of 0.88 was achieved using this dual-stream approach, along with enhanced predictive probabilities for all species thus improving the reliability of the predicted results. This approach reduces the manual processing bottleneck of individual tree segmentation and labelling and demonstrates the value of combining raw point-clouds and point-cloud metrics into a deep learning framework, offering a scalable and operational solution for reliable species predictions.

Climate change has significantly impacted the wildfire regimes in lodgepole pine forests, resulting in prolonged fire seasons and altered fire behaviour. In North America, fire patterns have shifted towards more frequent and severe wildfires after a century of fire suppression. In response, silviculture practices in fire-prone areas should aim to restore diverse forest structures that are resistant or resilient to wildfires. In Western Canada, where forestry is a key industry, interest in seeking silvicultural solutions for promoting forest resilience to wildfires has increased following the devastating wildfire seasons between 2017 and 2023. Irregular shelterwood, a silvicultural system with a relatively short history of implementation in British Columbia, has been deployed in ecologically sensitive areas to promote structural heterogeneity and meet management goals for biodiversity and wildlife values. Although the impacts of irregular shelterwood on wildlife habitat and abundance have been well studied, the interaction between wildfire and the stand structure created by irregular shelterwood remains poorly understood. To understand the effectiveness of the irregular shelterwood in building wildfire resilience, we present a study of a lodgepole pine stand that was treated with irregular shelterwood and partially burned in a wildfire in 2017. This study collected ground fuel, canopy fuel, and tree data from four stand types (irregular shelterwood treated-burnt, treated-unburnt, untreated-burnt, and untreated-unburnt) and analyzed the difference in char height and fire-induced mortality between burnt and unburnt conditions, with irregular shelterwood treatment being a variable. The results demonstrated reduced wildfire effect in the irregular shelterwood stand in this region of British Columbia. This observation was made at a stage where the openings have not been colonized by regeneration. This case study provides valuable insights into the effectiveness of irregular shelterwood in mitigating wildfire risk, and proposes a potential silviculture solution to promote forest resilience to wildfire.

Airborne laser scanning (ALS) data is increasingly being used for accurate estimations of forest inventory attributes, such as tree height and volume. However, determining tree species information, an important attribute required by inventories for a range of forest management applications, is challenging to extract from ALS data. This is due to complex species assemblages and vertical structures in certain forest environments, and a lack of spectral information for most ALS sensors. Our objective was to estimate tree species compositions in a Canadian boreal forest environment using ALS data and point-based deep learning techniques. To accomplish this, we utilised existing polygonal forest resource information to develop a large comprehensive dataset of tree species compositions across a 630,000 ha forest management area. We then used an adapted PointAugment generative adversarial network (GAN) coupled with the dynamic graph convolutional neural network to estimate tree species proportions. This innovative deep learning approach resulted in an overall R2adj of 0.61 for all tree species proportions. A weighted F1 score of 63% was observed when classifying the leading species of a plot, 66% for leading genus, and 85% when distinguishing between coniferous and deciduous dominated plots. Furthermore, the PointAugment GAN successfully generated augmented point clouds of forest plots which can alleviate issues associated with limited training samples for use in deep learning. This approach demonstrates the capability of point-based deep learning techniques to accurately estimate tree species compositions from ALS point clouds within an expansive forested region, characterized by a complex management history and a diversity of tree species. The source code along with the pretrained models are available at https://github.com/Brent-Murray/point-dl/tree/main/Pytorch/models/PointAugment.

We assessed the impacts of three approaches to thinning from below with varying spatial patterns on several stand and individual tree variables for interior Douglas-fir ( Pseudotsuga menziesii var. glauca (Beissn.) Franco), interior spruce ( Picea glauca (Moench) Voss × Picea engelmannii Engelm.), and lodgepole pine ( Pinus contorta Dougl. Ex Loud. var. latifolia Englem.) in central British Columbia, Canada. The three thinning treatments were two experimental “clumped” treatments (3 m Clumped and 5 m Clumped) and the Standard (more uniform spacing) thinning treatment that was employed operationally at that time. We used long-term data from 24 plots measured five times over 21 years. Thinning increased stand basal area increment, with the plots that received the 5 m Clumped treatment having significantly higher periodic annual relative basal area increment than the unthinned Control plots. The responses for the two clumped treatments were not any lower than the Standard. The 3 m Clumped treatment was best if one is concerned about fast recovery of the growing space; however, the 5 m Clumped spacing treatment may be preferable if higher individual tree vigour is needed for resistance and resilience to fire, insects, and disease.

Western redcedar has high economic value and has been traditionally used for cultural purposes by Indigenous communities. Second-growth redcedar is potentially growing faster due to lower planting densities, fertilization, and tree breeding. Little quantitative information is available about the impact of management practices on wood quality, particularly heartwood extractives. This study evaluated the effects of growth rate and site on heartwood extractives at two locations aged 70-90 years in British Columbia, Canada. A three-parameter sigmoid model was fit to the data using nonlinear mixed effects to analyze the relationship between heartwood extractives relative to cambial age, growth rate, and sampling site. The southern site had significantly higher cumulative extractive concentrations, while all extractive concentrations increased faster. This study shows that smaller trees will reach their peak concentrations earlier than larger trees. Results show that faster growth through active management of western redcedar may lead to increased and more uniformly distributed content of heartwood extractives.

Understanding the dynamics of treeline ecotones under global change requires long-term ecological and environmental data. The Stillberg ecological treeline research site in the Swiss Alps was established in 1975 by planting 92,000 seedlings of Larix decidua, Pinus cembra and Pinus mugo ssp. uncinata, and has been continuously monitored since then. Here, we present curated long-term data acquired over almost 50 years at the Stillberg site, and we synthesise the major research findings. The long-term datasets comprise 6.5 million ecological and environmental records from the 40-year afforestation experiment, as well as from a 9-year free-air CO2 enrichment experiment crossed with a 6-year soil warming experiment, a 12-year nutrient addition experiment and an 8-year multifactorial tree seedling recruitment experiment. Our datasets further include 38 million records of 25 meteorological parameters measured at an hourly resolution from 1975 to 1996, and at a 10 min resolution since 1997. We provide all datasets and the corresponding metadata as open research data. Almost five decades of research in this treeline ecotone showed high mortality after tree establishment that was closely related to microclimatic variability. The two Pinus species survived at a much lower rate than L. decidua, due to indirect pathogen interactions. Furthermore, CO2 enrichment only increased growth of L. decidua, while warming increased growth of P. mugo ssp. uncinata and two Vaccinium shrub species. Enhanced nutrient availability stimulated growth in tree and understorey shrub species. In addition, soil warming and CO2 enrichment stimulated microbial activity and decreased soil carbon stocks. These findings improve our understanding of ecological processes in the treeline ecotone under global change and confirm the importance of tree growth and establishment limitations. The enhanced availability and quality of these long-term data are expected to foster whole-system approaches and transdisciplinary research syntheses, supporting the development of effective global change adaptation strategies.

Drought can impact forests directly causing a decrease of growth, but also increase the vulnerability of trees to secondary pests and pathogens, causing additional loss of volume production. To develop new silvicultural strategies, it is crucial to understand if thinning can promote resilience of the remaining trees to drought by enhancing an efficient use of resources. Given projected drier vegetation periods in Southern Sweden, the aim of the study was to determine how tree growth is affected by severe summer droughts under different thinning regimes. We used an experiment established in 1991 in a 40-year-old pure oak (Quercus robur L.) stand with two thinning intensities and an unthinned control. We collected tree cores before and after specific drought events occurring after treatment. We observed that heavy thinning intensity increased drought resistance, and decreased recovery time and growth reduction when the time since the last intervention was 4-5 years. Our results suggest that heavy and frequent thinning interventions would be an appropriate management alternative to alleviate drought stress in pure oak stands close to the northern edge of their distribution.

Laser scanning sensors mounted on drones enable on-demand quantification of forest structure through the collection of high-density point clouds (500+ points m−2). These point clouds facilitate the detection of individual trees enabling the quantification of growth-related variables within a stand that can inform precision management. We present a methodology to link incremental growth data obtained from tree cores with crown models derived from drone laser scanning, quantifying the relative growth condition of individual trees and their neighbours. We stem-mapped 815 trees across five stands in north-central British Columbia, Canada of which 16% were cored to quantify recent basal area growth. Point clouds from drone laser scanning and orthomosaic imagery were used to locate trees, model three-dimensional crown features, and derive competition metrics describing the relative distribution of crown sizes. Local access to water and light were simulated using topographic wetness and potential solar irradiance indices derived from high-resolution terrain and surface models. Wall-to-wall predictions of recent basal area growth were produced from the best-performing model and summarized across a grid alongside a tree-level competition index. Overall, crown volume was most strongly correlated with observed differences in 5-year basal area increment (R2 = 0.70, P < .001). Competition and solar irradiance metrics were significant as univariate predictors (P < .001) but nonsignificant when included in multivariate models with crown volume. Using predictions from the best-performing model and laser-scanning-derived competition metrics, we present a newly developed growth competition index to assess variability and inform commercial thinning prescription prioritization. Growth predictions, competition metrics, and the growth competition index are summarized into maps that could be used in an operational workflow. Our methodology presents a new capacity to capture and quantify intra-stand variation in growth by combining competition metrics and measures of recent growth with high-density drone laser scanning data.

Alpine treeline ecosystems are generally expected to advance with increasing temperatures and after land-use abandonment. Multiple interacting factors modify this trend. Understanding the long-term processes underlying treeline advance is essential to predict future changes in structure and function of mountain ecosystems. In a valley in the Central Swiss Alps, we re-assessed a 40-year-old survey of all treeline trees (>0.5 m height) and disentangled climate, topographical, biotic, and disturbance (land use and avalanche risk) factors that have led to treeline advance with a combination of ground-based mapping, decision tree, and dendroecological analyses. Between the first ground survey in 1972/73 and the resurvey in 2012, treeline advanced on average by 10 meters per decade with a maximum local advance of 42 meters per decade. Larch consistently advanced more on south-facing slopes, while pine advance was greater on north-facing slopes. Newly established spruce mostly represented infilling below the previous treeline. The forefront of treeline advance above 2330 m a.s.l. occurred mainly on favorable microsites without competing dwarf shrub vegetation. At slightly lower elevations, treeline advanced mainly on sites that were used for agriculture at the beginning of the 20th century. This study indicates that although treeline advances under the effect of climate warming, a combination of additional ecological factors controls this advance at regional and local scales.

Understanding the spatial patterns of trees and their interactions can reveal the ecological processes driving forest stand structure and stand development over time. We assessed temporal changes in tree spatial patterns in uneven-aged interior Douglas-fir (Pseudotsuga menziesii var. glauca (Beissn.) Franco) dominated stands in central British Columbia, Canada. Data were available on 24 plots in three blocks over 21 years, 18 of which had received pre-commercial thinning (PCT) treatments of varying intensity. We first applied the Clark and Evans aggregation index and the L function, a transformation of Ripley’s K function, to describe the spatial pattern of live trees in thinned and unthinned plots over time. Second, we analysed the spatial correlations between live tree diameters using the mark correlation function and the mark variogram. Third, the spatial pattern of dead trees in the unthinned plots after 21 years was analysed. Lastly, we tested the spatial relationship between dead and surviving trees in the unthinned plots. In all three blocks, the spatial patterns of live trees in the unthinned plots were clustered through time. The moderate thinning treatments had random or regular spatial patterns that remained unchanged through time, because of reduced mortality rates and low levels of ingrowth. The heavier thinning initially retained a clustered pattern at small inter-tree distances; however, this changed to a random pattern with increasing distance 15 years post-thinning and remained as such until the end of the study period. Tree diameters were not spatially autocorrelated in the thinned plots, although there was positive spatial correlation of tree diameters in the unthinned plots, probably due to competitive growth inhibition among neighbouring trees. Dead trees were primarily smaller in size and were significantly clustered at all spatial scales. The lack of spatial relationship between dead and surviving trees indicated that mortality was a random process. Our study contributes to a better understanding of how spatial patterns of trees change over time in these stands, which could help in the design of silvicultural regimes that mimic natural processes.

How seedling mortality and browsing affects species composition of regenerating forests has been mostly studied on a small scale. Yet, large-scale analyses based on extensive data are essential for robust prediction of species composition in young forests. In this study, we used a dataset from a national inventory of young forests (1-4 metres in height) to investigate the species composition of young forests across Sweden. We found that most of the regenerated forest area (almost 90%) was planted with Norway spruce (southern Sweden) and Scots pine (northern Sweden). Regeneration of Norway spruce was generally relatively successful but as a consequence of seedling mortality and browsing, almost 40% of the area regenerated with Scots pine will probably not develop into pine-dominated stands. Thus, low survival of Scots pine seedlings and trees can profoundly change the trajectory of species composition of the young forest from what was originally intended, and a large proportion of the young stands may develop into mixtures of conifers and broadleaves. While such mixtures may benefit certain biodiversity and ecosystem services, a loss of Scots pine dominated stands may also have adverse impacts on the economic returns as well as pine-dependent biodiversity and recreational values.

The increasing disturbances in monocultures around the world are testimony to their instability under global change. Many studies have claimed that temporal stability of productivity increases with species richness, although the ecological fundamentals have mainly been investigated through diversity experiments. To adequately manage forest ecosystems, it is necessary to have a comprehensive understanding of the effect of mixing species on the temporal stability of productivity and the way in which it is influenced by climate conditions across large geographical areas. Here, we used a unique dataset of 261 stands combining pure and two-species mixtures of four relevant tree species over a wide range of climate conditions in Europe to examine the effect of species mixing on the level and temporal stability of productivity. Structural equation modelling was employed to further explore the direct and indirect influence of climate, overyielding, species asynchrony and additive effect (i.e. temporal stability expected from the species growth in monospecific stands) on temporal stability in mixed forests. We showed that by adding only one tree species to monocultures, the level (overyielding: +6%) and stability (temporal stability: +12%) of stand growth increased significantly. We identified the key effect of temperature on destabilizing stand growth, which may be mitigated by mixing species. We further confirmed asynchrony as the main driver of temporal stability in mixed stands, through both the additive effect and species interactions, which modify between-species asynchrony in mixtures in comparison to monocultures. Synthesis and applications. This study highlights the emergent properties associated with mixing two species, which result in resource efficient and temporally stable production systems. We reveal the negative impact of mean temperature on temporal stability of forest productivity and how the stabilizing effect of mixing two species can counterbalance this impact. The overyielding and temporal stability of growth addressed in this paper are essential for ecosystem services closely linked with the level and rhythm of forest growth. Our results underline that mixing two species can be a realistic and effective nature-based climate solution, which could contribute towards meeting EU climate target policies.